Even in the past, broad social categories like status played a crucial role in language variation and change. However, language use is also linked to the existence of more locally defined networks of people, and for this reason the notion of community of practice (CofP), i.e. groups of individuals who come together around particular activities, has recently come to the fore in the study of language history. However, the interplay between locally negotiated practices and broader social structure in language history remains unclear to date. Focusing on the context of late medieval Florence, my project represents the first attempt to combine the CofP framework with insights provided by more traditional categories. Fifteenth-century Florentine was characterised by a rich phonological and morphosyntactic variation that remains largely unaccounted for from a sociolinguistic perspective. Aiming to shed light on the social meaning of this variation, I analyse a hitherto untapped linguistic source, the 1427 Catasto, a survey of households for tax purposes that records the language of a significant part of the Florentine population and is rich with demographic information on individual writers. Through quantitative and qualitative analysis, the project assesses the role of status, of CofPs that formed around districts and guilds, and the interplay between the two dimensions, aiming to bridge the gap between the macro- and micro-level in our understanding of sociolinguistic variation.